Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[release-1.24] Backports for 2022-10 #6227

Merged
merged 14 commits into from
Oct 7, 2022

Conversation

brandond
Copy link
Member

@brandond brandond commented Oct 6, 2022

Proposed Changes

Types of Changes

Backports

Verification

See linked issues

Testing

Linked Issues

User-Facing Change

The embedded metrics-server version has been bumped to v0.6.1
The ServiceLB (klipper-lb) service controller is now integrated into the K3s stub cloud controller manager.
Events recorded to the cluster by embedded controllers are now properly formatted in the service logs.
Fixed an issue with the apiserver network proxy that caused `kubectl exec` to occasionally fail with `error dialing backend: EOF`
Fixed an issue with the apiserver network proxy that caused `kubectl exec` and `kubectl logs` to fail when a custom kubelet port was used, and the custom port was blocked by firewall or security group rules.
The embedded Traefik version has been bumped to v2.9.1 / chart 12.0.0

Further Comments

Signed-off-by: Brad Davidson <[email protected]>
... and drop legacy ClusterID support.

Signed-off-by: Brad Davidson <[email protected]>
@brandond brandond marked this pull request as ready for review October 6, 2022 22:51
@brandond brandond requested a review from a team as a code owner October 6, 2022 22:51
brandond and others added 10 commits October 6, 2022 23:08
If CCM and ServiceLB are both disabled, don't run the cloud-controller-manager at all;
this should provide the same CLI flag behavior as previous releases, and not create
problems when users disable the CCM but still want ServiceLB.

Signed-off-by: Brad Davidson <[email protected]>
We should be reading from the hijacked bufio.ReaderWriter instead of
directly from the net.Conn. There is a race condition where the
underlying http handler may consume bytes from the hijacked request
stream, if it comes in the same packet as the CONNECT header. These
bytes are left in the buffered reader, which we were not using. This was
causing us to occasionally drop a few bytes from the start of the
tunneled connection's client data stream.

Signed-off-by: Brad Davidson <[email protected]>
The kubelet port can be overridden by users; we shouldn't assume its always 10250

Signed-off-by: Brad Davidson <[email protected]>
CA cert will never be equal to the serving-kube-apiserver cert so it seems like a copy-paste error.

Signed-off-by: Vladimir Pouzanov <[email protected]>
Replace ETCD-JOIN-STABLE-SECOND with ETCD-JOIN-LATEST-FIRST. We don't
support joining down-level servers to existing clusters, as the new
down-level server will try to deploy older versions of the packaged
manifests.

Signed-off-by: Brad Davidson <[email protected]>
@codecov-commenter
Copy link

codecov-commenter commented Oct 7, 2022

Codecov Report

Base: 9.71% // Head: 9.73% // Increases project coverage by +0.01% 🎉

Coverage data is based on head (625d961) compared to base (205ccc7).
Patch coverage: 1.73% of modified lines in pull request are covered.

Additional details and impacted files
@@               Coverage Diff               @@
##           release-1.24   #6227      +/-   ##
===============================================
+ Coverage          9.71%   9.73%   +0.01%     
===============================================
  Files               138     140       +2     
  Lines             10064   10130      +66     
===============================================
+ Hits                978     986       +8     
- Misses             8887    8942      +55     
- Partials            199     202       +3     
Flag Coverage Δ
unittests 9.73% <1.73%> (+0.01%) ⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files Coverage Δ
pkg/agent/tunnel/tunnel.go 0.00% <0.00%> (ø)
pkg/cli/server/server.go 0.00% <0.00%> (ø)
pkg/cloudprovider/cloudprovider.go 0.00% <0.00%> (ø)
pkg/cloudprovider/instances.go 0.00% <0.00%> (ø)
pkg/cloudprovider/loadbalancer.go 0.00% <0.00%> (ø)
pkg/cloudprovider/servicelb.go 0.00% <0.00%> (ø)
pkg/daemons/agent/agent_linux.go 0.00% <0.00%> (ø)
pkg/daemons/config/types.go 64.91% <ø> (ø)
pkg/daemons/control/proxy/proxy.go 0.00% <ø> (ø)
pkg/daemons/control/server.go 0.00% <0.00%> (ø)
... and 10 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

Taint the first node so that the helm job doesn't run on it. In a real cluster the helm job would eventually succeed once all the servers were upgraded and had the new chart tarball.

Signed-off-by: Brad Davidson <[email protected]>
@brandond
Copy link
Member Author

brandond commented Oct 7, 2022

Drone is not currently syncing PRs, but other tests are green and a local CI run succeeds. Merging.

@brandond brandond merged commit c1c7b95 into k3s-io:release-1.24 Oct 7, 2022
@brandond brandond deleted the backports_release-1.24 branch June 6, 2024 21:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants